WO2010021108A1 - 補間フィルタリング方法、画像符号化方法、画像復号方法、補間フィルタリング装置、プログラム、及び集積回路 - Google Patents
補間フィルタリング方法、画像符号化方法、画像復号方法、補間フィルタリング装置、プログラム、及び集積回路 Download PDFInfo
- Publication number
- WO2010021108A1 WO2010021108A1 PCT/JP2009/003886 JP2009003886W WO2010021108A1 WO 2010021108 A1 WO2010021108 A1 WO 2010021108A1 JP 2009003886 W JP2009003886 W JP 2009003886W WO 2010021108 A1 WO2010021108 A1 WO 2010021108A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- pixel
- block
- interpolation
- weighting factor
- pixels
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/523—Motion estimation or motion compensation with sub-pixel accuracy
Definitions
- the present invention relates to video encoding and decoding, and more particularly to a method of interpolating pixel values in a block-unit video codec using decimal pixel precision motion compensated prediction.
- video coding standards have been established. Such video standards include H.264. ITU-T recommendations that define the 26 system, ISO / IEC standards that define the MPEG system, and the like are included. Currently, the latest and highest level video coding standard is H.264. H.264 / MPEG-4AVC.
- All of these standards are based on hybrid video coding consisting of the following main processes.
- A Each video frame is divided into two-dimensional pixel blocks in order to perform data compression of each video frame at a block level.
- the spatio-temporal video information is decorrelated by applying a spatio-temporal prediction method to each block, converting a prediction error from the spatial domain to the frequency domain, and calculating a coefficient.
- C The total amount of data is reduced by quantizing the calculated coefficients.
- D The quantized coefficient and the prediction parameter are encoded by an entropy encoding algorithm to compress the data.
- the differential pulse code modulation (DPCM) system that simply transmits the difference between the block of the input video sequence and the prediction based on the already encoded block. Is adopted.
- One of the prediction methods that can be adopted in the three video coding standards is motion compensation prediction.
- this prediction method at least one motion vector is determined for each block of video data in order to indicate image displacement accompanying movement of the subject and / or camera. Based on the determined motion vector, one block of image content can be predicted at least to some extent from the already encoded image content of the block. The difference between the predicted image content and the actual input image content is referred to as a prediction error, and this difference is encoded with the motion vector rather than the actual input image content.
- the amount of information to be encoded can be substantially reduced for most “natural” video sequences.
- FIG. 1 is a block schematic diagram showing a conventional image encoding device 100 compliant with the H.264 / AVC standard.
- This image encoding apparatus 100 is a prediction error signal that is a difference between a current block of a video (input signal) and a prediction signal of a current block based on a block (local decoded image) that has already been encoded and stored in the memory 140. Is provided.
- the memory 140 functions as a delay unit that makes it possible to compare the current signal value and the predicted signal generated from the previous signal value.
- the transform / quantization unit 120 transforms the calculated prediction error signal from the spatial domain to the frequency domain, and quantizes the obtained quantization coefficient.
- the entropy encoding unit 190 entropy encodes the quantized coefficient.
- the local decoded image is created by a decoding unit incorporated in the image encoding device 100.
- the decoding unit performs the encoding process in reverse.
- the inverse quantization / inverse transform unit 130 inversely quantizes the quantized coefficient and performs inverse transform on the inversely quantized coefficient.
- the adder 135 adds the decoded quantized prediction error signal to the prediction signal, and generates a reconstructed signal.
- the deblocking filter 137 may be applied to each reconstructed macroblock.
- the deblocking filter 137 is applied to the reconstructed signal that is the total value of the prediction signal and the quantized prediction error signal.
- the deblocked signal is a decoded signal, which is generally displayed.
- the H.264 / AVC deblocking filter 137 has a local adaptation function. When the blocking noise is large, an intensity low-pass filter is applied, and when the blocking noise is small, a low-intensity low-pass filter is applied.
- the strength of the deblocking filter 137 is determined by the so-called boundary strength bS and the gradient of pixel values sandwiching the boundary.
- the boundary strength bS is a parameter set as shown in Table 1 below according to the method of encoding the block p and the block q on both sides of the boundary.
- a linear deblocking filter using four coefficients is installed. This filter is applied to six adjacent input samples p 2 , p 1 , p 0 , q 0 , q 1 , q 2 . At this time, the boundary is located between p 0 and q 0 .
- p 0, new (p 2- (p 1 ⁇ 1) + (p 0 + q 0 + 1) >> 1) >> 1
- q 0, new (q 2- (q 1 ⁇ 1) + (q 0 + p 0 + 1) >> 1) >> 1
- the type of prediction used in the image encoding apparatus 100 is determined depending on whether the macroblock is encoded in the “intra” mode or in the “inter” mode.
- “intra” mode the video coding standard H.264 is used.
- H.264 / AVC uses a prediction method based on already encoded macroblocks of the same image to predict subsequent macroblocks.
- “inter” mode motion compensation prediction is performed between corresponding blocks of several consecutive frames.
- I-type images are error tolerant to encoded video sequences.
- an entry point to the bit stream of encoded data is provided by the I-type image in order to allow random access, i.e. to access an I-type image in the sequence of encoded video.
- the Switching between the intra mode, that is, the processing by the in-screen frame prediction unit 150, and the inter mode, that is, the processing by the motion compensation prediction unit 160, is controlled by a switch 180 (intra / inter switching unit).
- the macroblock predicts from the corresponding block of the previous frame using motion compensated prediction. Prediction is performed by the motion prediction unit 170 by receiving the current input signal and the locally decoded image. Motion prediction generates a two-dimensional motion vector that indicates the displacement between the current block and the corresponding block of an already decoded frame. Based on the predicted motion, the motion compensation prediction unit 160 transmits a prediction signal.
- the motion vector may be obtained with sub-pixel accuracy such as 1/2 pixel accuracy or 1/4 pixel accuracy.
- the sub-pixel precision motion vector may indicate a position in a frame where there is no valid pixel value that has already been decoded, that is, a sub-pixel position. Therefore, in order to perform motion compensation, spatial interpolation of pixel values is required. This is done by the interpolation filter 162.
- H. According to the H.264 / AVC standard, in order to obtain the pixel value at the sub-pixel position, a 6-tap Wiener interpolation filter with a fixed filter coefficient and a bilinear filter are applied.
- an adaptive interpolation filter may be used.
- the non-separable two-dimensional interpolation filter known from Patent Document 1 cited as a reference document is calculated separately for each frame by minimizing the prediction error energy.
- a set of filter coefficients is individually determined on the encoder side, and is transmitted as auxiliary information to the image decoding apparatus as indicated by a broken line in FIG. 1A.
- the transform / quantization unit 120 transforms and quantizes the prediction error signal that is the difference between the current input signal and the prediction signal, thereby quantizing. A coefficient is obtained.
- orthogonal transformation such as two-dimensional discrete cosine transformation (DCT) or an integer version thereof is performed.
- the reason for quantizing the coefficients is to reduce the amount of data to be encoded.
- the quantization process is controlled by a quantization table that indicates the accuracy and the number of bits used to encode each frequency coefficient with that accuracy.
- the low frequency component is more important than the high frequency component, so that the number of bits used for encoding the low frequency component is larger than the encoding of the high frequency component.
- the two-dimensional array of quantized coefficients After quantization, in order to pass the quantized coefficients to the entropy coding unit 190, the two-dimensional array of quantized coefficients must be converted into a one-dimensional column. This conversion is performed by scanning the array in a predetermined sequence.
- the entropy encoding unit 190 compresses and encodes the one-dimensional sequence of the quantized coefficients thus obtained by variable length encoding (VLC).
- VLC variable length encoding
- FIG. 2A is a schematic block diagram illustrating a structure of a corresponding image decoding apparatus.
- the entropy decoding unit 191 reversely performs entropy coding of the coefficient and motion information. This process also includes reverse scanning for converting the decoded sequence of coefficients into a two-dimensional data array necessary for reverse conversion. Thereafter, the decoded block of quantized coefficients is passed to the inverse quantization / inverse transform unit 121 and the decoded motion information is sent to the motion compensation prediction unit 160.
- pixel value interpolation may be required to perform motion compensation prediction.
- This interpolation is performed by the interpolation filter 162.
- the filter coefficient is also decoded by the entropy decoding unit 191 and supplied to the interpolation filter 162 as shown by the broken line in FIG. 2A.
- the result of the inverse transform is a quantized prediction error signal in the spatial domain, and is added by the adder 135 to the prediction signal transmitted from the inter-mode motion compensation prediction unit 160 or the intra-mode frame prediction unit 150.
- the reconstructed image may pass through the deblocking filter 137, and the obtained decoded signal is stored in the memory 140 and then passed to the intra-frame prediction unit 150 or the motion compensation prediction unit 160.
- an object of the present invention is to provide an improved method and apparatus for video encoding and decoding. Furthermore, another object of the present invention is to provide an improved method of motion compensated prediction and an improved method of pixel value interpolation required for performing motion compensated prediction with decimal pixel accuracy.
- the interpolation filtering method interpolates pixels of an image encoded in units of blocks. Specifically, a pixel selection step for selecting a plurality of pixels used for interpolation of the interpolation target pixel from the image, and a weighting factor determination for determining a weighting factor corresponding to each of the plurality of pixels selected in the pixel selection step And an interpolation step of calculating a pixel value of the interpolation target pixel by weighting and adding pixel values of the plurality of pixels using the weighting factor determined in the weighting factor determination step.
- the weighting factor determination step when any one of the plurality of pixels belongs to a peripheral block different from the target block to which the interpolation target pixel belongs, the pixel belongs to the target block. Reduce the pixel weighting factor.
- the weight of a pixel belonging to the same block as the interpolation target pixel can be relatively increased, and the weight of a pixel belonging to a block different from the interpolation target pixel can be relatively decreased.
- the influence of distortion resulting from encoding for each block is minimized.
- a more accurate prediction block can be generated.
- a distance from the interpolation target pixel to the boundary of the target block is calculated. Then, for each of the plurality of pixels, the calculated distance is used to determine whether a block boundary of the target block exists between the interpolation target pixel and the pixel. It may be determined whether it belongs to or not.
- the weighting factor of the pixels belonging to the peripheral block may be reduced as the boundary strength of the block boundary between the target block and the peripheral block increases. For example, when the block distortion between the target block and the peripheral block is large, an extremely small value may be set for the weight coefficient of the pixel belonging to the peripheral pixel. Thus, if the weighting factor is determined based on the characteristics of the block boundary, further improvement in the accuracy of the prediction block can be expected.
- both the target block and the peripheral block are intra-screen predictively encoded.
- the weight coefficient of the pixels belonging to the peripheral block may be reduced.
- the weighting factor of pixels belonging to the peripheral block may be made smaller than when the block boundary is not a macroblock boundary.
- the peripheral block when at least one of the target block and the peripheral block includes an encoding coefficient, the peripheral block is more than when the target block and the peripheral block do not include an encoding coefficient.
- the weighting factor of pixels belonging to may be reduced.
- the weighting factor determination step when at least one of the reference frame, the reference frame number, and the motion vector value of the target block and the peripheral block is different, the reference frame and the reference frame number of the target block and the peripheral block , And the case where the motion vectors are all the same, the weight coefficients of the pixels belonging to the peripheral block may be made smaller.
- the weighting factor of the pixel belonging to the peripheral block is increased as the gradient of the pixel value of two pixels sandwiching the boundary between the target block and the peripheral block is larger. It may be small. This can also be expected to improve the accuracy of the prediction block.
- the image encoding method sequentially encodes a plurality of encoding target blocks.
- the method includes a prediction block generation step for generating a prediction block of the encoding target block, and an encoding step for encoding the encoding target block using the prediction block.
- the prediction block generation step includes: a pixel selection step for selecting a plurality of pixels used for interpolation of the interpolation target pixel from the image; and a weight for determining a weighting factor corresponding to each of the plurality of pixels selected in the pixel selection step.
- a coefficient determination step an interpolation step for calculating a pixel value of the interpolation target pixel by weighting and adding pixel values of the plurality of pixels using the weighting factor determined in the weighting factor determination step, and the pixel interpolating A prediction step of generating the prediction block based on the processed image.
- the weight coefficient determination step when the pixel belongs to a peripheral block different from the target block to which the interpolation target pixel belongs, the weight coefficient of the pixel is made smaller than when the pixel belongs to the target block. According to the above configuration, a highly accurate prediction block can be generated, so that the encoding efficiency of the encoding target block can be improved.
- the weighting factor may be encoded. Thereby, the step of calculating the weighting factor on the decoder side can be omitted.
- the image decoding method generates a decoded block by decoding an encoded signal. Specifically, a prediction block generation step for generating a prediction block of the decoded block and a decoding step for decoding the encoded signal using the prediction block are included.
- the prediction block generation step includes: a pixel selection step for selecting a plurality of pixels used for interpolation of the interpolation target pixel from the image; and a weight for determining a weighting factor corresponding to each of the plurality of pixels selected in the pixel selection step.
- a coefficient determination step an interpolation step for calculating a pixel value of the interpolation target pixel by weighting and adding pixel values of the plurality of pixels using the weighting factor determined in the weighting factor determination step, and the pixel interpolating A prediction step of generating the prediction block based on the processed image.
- the weight coefficient determination step when the pixel belongs to a peripheral block different from the target block to which the interpolation target pixel belongs, the weight coefficient of the pixel is made smaller than when the pixel belongs to the target block.
- the weighting factor is calculated on the decoder side based on the same rules as those on the encoder side as in the above configuration, it is not necessary to encode the weighting factor. Thereby, the encoding efficiency of the encoding target block can be further improved.
- the interpolation filtering apparatus interpolates pixels of an image encoded in units of blocks. Specifically, a pixel selection unit that selects a plurality of pixels used for interpolation of the interpolation target pixel from the image, and a weighting factor determination that determines a weighting factor corresponding to each of the plurality of pixels selected by the pixel selection unit And an interpolation unit that calculates the pixel value of the interpolation target pixel by weighting and adding the pixel values of the plurality of pixels using the weighting factor determined by the weighting factor determination unit.
- the weighting factor determination unit is configured such that when any of the plurality of pixels belongs to a peripheral block different from the target block to which the interpolation target pixel belongs, the weighting factor of the pixel is greater than when the pixel belongs to the target block. Make it smaller.
- the program according to an aspect of the present invention causes a computer to interpolate pixels of an image encoded in units of blocks. Specifically, a pixel selection step for selecting a plurality of pixels used for interpolation of the interpolation target pixel from the image, and a weighting factor determination for determining a weighting factor corresponding to each of the plurality of pixels selected in the pixel selection step And an interpolation step of calculating a pixel value of the interpolation target pixel by weighting and adding pixel values of the plurality of pixels using the weighting factor determined in the weighting factor determination step.
- the weighting factor determination step when any of the plurality of pixels belongs to a peripheral block different from the target block to which the interpolation target pixel belongs, the weighting factor of the pixel is higher than when the pixel belongs to the target block. Make it smaller.
- the integrated circuit interpolates pixels of an image encoded in units of blocks. Specifically, a pixel selection unit that selects a plurality of pixels used for interpolation of the interpolation target pixel from the image, and a weighting factor determination that determines a weighting factor corresponding to each of the plurality of pixels selected by the pixel selection unit And an interpolation unit that calculates the pixel value of the interpolation target pixel by weighting and adding the pixel values of the plurality of pixels using the weighting factor determined by the weighting factor determination unit.
- the weighting factor determination unit is configured such that when any of the plurality of pixels belongs to a peripheral block different from the target block to which the interpolation target pixel belongs, the weighting factor of the pixel is greater than when the pixel belongs to the target block. Make it smaller.
- the present invention can be realized not only as an interpolation filtering method (device) but also as an integrated circuit that realizes these functions, or as a program that causes a computer to execute such functions. Needless to say, such a program can be distributed via a recording medium such as a CD-ROM and a transmission medium such as the Internet.
- an interpolation filter for interpolating a pixel value at a specific position is selected based on the distance from the specific position to the adjacent block boundary.
- a method for interpolating pixel values in a block-by-block video encoder or video decoder using motion compensated prediction with decimal pixel accuracy includes a step of applying one interpolation filter selected from a plurality of interpolation filters to a pixel value to be interpolated, and the interpolation filter extends from an image position from which an interpolation value is obtained to the nearest neighboring block boundary in the vertical direction.
- the interpolation filter extends from an image position from which an interpolation value is obtained to the nearest neighboring block boundary in the vertical direction.
- the method of the present invention further includes a step of determining whether or not there is a block boundary at a predetermined horizontal distance and / or a predetermined vertical distance from the image position for which an interpolation value is determined.
- An interpolation filter is selected based on the result of. In this way, for example, an optimal interpolation filter may be selected depending on whether a horizontal or vertical block boundary is in the vicinity.
- the predetermined horizontal direction distance and the predetermined vertical direction distance are half of the number of taps in the horizontal direction and the vertical direction of the interpolation filter, respectively. Further, the predetermined horizontal direction distance and the predetermined vertical direction distance may be set to values smaller than one half of the number of taps of the interpolation filter in the horizontal direction and the vertical direction, respectively.
- an interpolation filter can be selected based on whether the filter support touches the block boundary, in other words, based on whether the filter tap intersects the block boundary.
- the method of the present invention further includes a step of obtaining a first horizontal distance from an image position from which an interpolation value is obtained to an adjacent vertical block boundary on the right side, and a left side from the image position from which the interpolation value is obtained.
- an optimal interpolation filter may be selected according to the distance to the surrounding block boundary.
- the interpolation filter may be selected based on a minimum value between the first horizontal direction distance and the second horizontal direction distance and a minimum value between the first vertical direction distance and the second vertical direction distance. This is particularly advantageous when the interpolation filter is bilaterally symmetric, for example when the left filter tap is weighted with the same coefficient as the filter tap at the right symmetric position.
- the interpolation filter further includes an encoding parameter indicating an encoding method of blocks on the left and right sides of the vertical block boundary and / or an encoding method of blocks above and below the horizontal block boundary. It is selected based on the encoding parameter shown. More preferably, the encoding parameter is such that each block is intra-coded, inter-coded, and / or whether a slice to which each block belongs is different and / or a macroblock to which each block belongs. Whether they are different and / or whether each block contains coding coefficients and / or whether there is a common reference frame for each block and / or each block has the same motion vector Indicates whether or not.
- the type of interpolation filter may be determined according to the type and strength of the deblocking filter and / or loop filter applied to a specific boundary of the reference frame. Different interpolation filters may be applied to block boundaries to which the high-intensity deblocking and / or loop filter is applied, and to block boundaries that are less filtered, or that are not filtered at all. Therefore, it is possible to select an optimal interpolation filter according to the type of the neighboring block boundary.
- the interpolation filter to be applied may be changed depending on, for example, whether the block boundary is a boundary between different slices, a macroblock boundary, or a block boundary in a macroblock.
- the coding parameter can also be an indicator of the degree of blocking artifacts that will appear at this particular boundary, so an optimal interpolation filter may be selected accordingly.
- the encoding parameter can also indicate which of the upper (left) block and the lower (right) block of the boundary is encoded more faithfully, so that the boundary can be improved to improve the prediction accuracy. Interpolation may be performed by paying attention to the pixel value on one side of.
- the interpolation filter is further selected based on a gradient between pixel values sandwiching a block boundary in the vertical direction and / or a gradient between pixel values sandwiching a block boundary in the horizontal direction.
- the gradient can also be some indication of the degree of blocking artifacts that appear at the block boundaries, so the optimal interpolation filter may be selected accordingly.
- block boundaries where blocking artifacts are prominent for example, it may be better not to consider pixels on the other side of the boundary at the time of calculating interpolated pixels, or to focus on pixels that are not adjacent to the boundary.
- a method for encoding and / or decoding a moving image sequence in order to encode or decode an image in units of blocks, each image of the moving image sequence is divided into a plurality of blocks, and one block of the plurality of blocks from the encoded or decoded blocks is divided. Applying the motion compensated prediction to obtain the predicted pixel value of the block using the pixel value interpolation method according to the first aspect of the present invention.
- the encoding method further includes adapting the plurality of interpolation filters to the moving image sequence by minimizing the prediction error energy.
- the prediction error energy is obtained by calculating the difference between the prediction block and the corresponding block of the moving image sequence to be encoded. In this way, it is possible to determine an optimal set of interpolation filters corresponding to the moving image sequence to be encoded. By performing motion compensation prediction based on an optimal interpolation filter, prediction accuracy can be improved, and as a result, overall coding efficiency is increased.
- an adaptation result is transmitted to the decoder side by encoding filter information indicating a plurality of adapted interpolation filters. Thereafter, the decoder side may decode the filter information indicating the plurality of interpolation filters, and then use the adapted interpolation filter.
- a video encoder and a video decoder configured to encode or decode a moving image sequence according to the second aspect of the present invention are provided.
- a computer program product comprising a computer readable medium having recorded thereon computer readable program code adapted to perform the method.
- the weight of a pixel belonging to the same block as the interpolation target pixel can be relatively increased, and the weight of a pixel belonging to a block different from the interpolation target pixel can be relatively decreased.
- the influence of distortion resulting from encoding for each block is minimized.
- a more accurate prediction block can be generated.
- FIG. 1A is a block diagram illustrating an image encoding device.
- FIG. 1B is a flowchart showing the operation of the image coding apparatus in FIG. 1A.
- FIG. 2A is a block diagram illustrating an image decoding apparatus.
- FIG. 2B is a flowchart showing the operation of the image decoding apparatus in FIG. 2A.
- FIG. 3A is a block diagram illustrating an interpolation filter.
- FIG. 3B is a flowchart showing the operation of the interpolation filter of FIG. 3A.
- FIG. 4 is a diagram illustrating a block of video data and an interpolation filter used for motion compensation prediction.
- FIG. 5 is a schematic diagram showing an interpolation filter selection method according to the embodiment of the present invention.
- FIG. 1B is a flowchart showing the operation of the image coding apparatus in FIG. 1A.
- FIG. 2A is a block diagram illustrating an image decoding apparatus.
- FIG. 2B is a flowchart showing the operation
- FIG. 6 is a schematic diagram illustrating an interpolation filter selection method according to another embodiment of the present invention.
- FIG. 7A is a diagram illustrating a state where the filter tap does not cross the block boundary.
- FIG. 7B is a diagram illustrating a state in which a part (one pixel) of the filter tap exceeds the block boundary.
- FIG. 7C is a diagram illustrating a state in which a part (two pixels) of the filter tap exceeds the block boundary.
- FIG. 7D is a diagram illustrating a state in which the interpolation target pixel is located at the block boundary.
- FIG. 7E is a view symmetrical with FIG. 7C.
- FIG. 7F is a view symmetrical with FIG. 7B.
- FIG. 7G is a diagram illustrating a state where the block boundary is exceeded at two locations of the filter tap.
- FIG. 1A is a block diagram of the image encoding device 100.
- FIG. 1B is a flowchart showing the operation of the image encoding device 100.
- the basic configuration of the image encoding device 100 is common to the conventional example and the present invention.
- the image encoding device 100 includes a subtractor 110, a transform / quantization unit 120, an inverse quantization / inverse transform unit 130, an adder 135, a deblocking filter 137, an entropy.
- An encoding unit 190 and a prediction block generation unit (not shown) are provided.
- the image encoding device 100 sequentially encodes a plurality of encoding target blocks constituting an encoding target image.
- the subtractor 110 subtracts the prediction block (prediction signal) from the encoding target block (input signal) to generate a prediction error signal.
- the transform / quantization unit 120 performs DCT transform (Discrete Cosine Transformation) and also quantizes the prediction error signal to generate a quantized coefficient.
- the entropy encoding unit 190 entropy encodes the quantized coefficient to generate an encoded signal. Note that entropy coding may be performed including the motion compensation data generated by the motion prediction unit 170, the filter coefficient of the interpolation filter 162, and the like together with the quantization coefficient.
- the inverse quantization / inverse transform unit 130 inversely quantizes the quantized coefficient and also performs inverse DCT transform to generate a quantized prediction error signal.
- the adder 135 adds the quantized prediction error signal and the prediction block to generate a reconstructed signal.
- the deblocking filter 137 removes block distortion from the reconstructed signal and generates a decoded signal.
- the prediction block generation unit generates a prediction block in which the encoding target block is predicted based on an image encoded before the encoding target block (input signal).
- the prediction block generation unit includes a memory 140, an in-frame frame prediction unit 150, a motion compensation prediction unit 160, an interpolation filter 162, a motion prediction unit 170, and a switch 180.
- the memory 140 functions as a delay device that temporarily stores the decoded signal. More specifically, the blocks quantized by the transform / quantization unit 120 and dequantized by the inverse quantization / inverse transform unit are sequentially stored, and one image (picture) is stored.
- the intra-frame prediction unit 150 performs intra-screen prediction on the decoded signal to generate a prediction signal.
- the interpolation filter 162 spatially interpolates the pixel value of the decoded signal prior to motion compensation prediction.
- the motion prediction unit 170 performs motion prediction based on the decoded signal and the next block to be encoded, and generates motion data (motion vector).
- the motion compensation prediction unit 160 performs motion compensation prediction based on the decoded signal and the motion data, and generates a prediction block.
- the switch 180 selects either the “intra” mode or the “inter” mode as the prediction mode.
- the prediction block output from the switch 180 is a signal obtained by predicting the next encoding target block.
- the subtractor 110 subtracts the prediction block from the encoding target block to generate a prediction error signal (S11).
- the transform / quantization unit 120 performs DCT transform on the prediction error signal and quantizes it to generate a quantized coefficient (S12).
- the entropy coding unit 190 entropy codes the quantization coefficient, the motion data, and the filter coefficient to generate a coded signal (S13).
- the inverse quantization / inverse transform unit 130 inversely quantizes the quantized coefficient and also performs DCT inverse transform to generate a quantized prediction error signal (S14). .
- the adder 135 adds the quantized prediction error signal and the prediction block to generate a reconstructed signal.
- the deblocking filter 137 removes block distortion from the reconstructed signal and generates a decoded signal.
- the prediction block generation unit generates a prediction block based on the decoded signal (S15). Details of the prediction block generation processing will be described later.
- FIG. 2A is a block diagram of the image decoding apparatus 101.
- FIG. 2B is a flowchart showing the operation of the image decoding apparatus 101. Note that the basic configuration of the image decoding apparatus 101 is common to the conventional example and the present invention.
- the image decoding apparatus 101 includes an entropy decoding unit 191, an inverse quantization / inverse conversion unit 121, an adder 135, a deblocking filter 137, a post filter 139, and a prediction block generation unit. (Not shown).
- the image decoding apparatus 101 decodes the encoded signal encoded by the image encoding apparatus 100 shown in FIG. 1A to generate a decoded block (decoded signal). Note that functional blocks common to the image encoding device 100 shown in FIG. 1A are assigned the same reference numerals, and detailed descriptions thereof are omitted.
- the entropy decoding unit 191 performs entropy decoding on the encoded signal (input signal) output from the image encoding device 100, and obtains quantization coefficients, motion data, and filter coefficients.
- the post filter 139 is, for example, a Wiener filter applied to the decoded signal, and improves the subjective image quality of the image.
- the prediction block generation unit includes a memory 140, an in-screen frame prediction unit 150, a motion compensation prediction unit 160, an interpolation filter 162, and a switch 180.
- the prediction block generation unit has the same basic configuration and operation, but omits the motion prediction unit 170 and acquires motion data from the entropy decoding unit 191 and the filter coefficient of the interpolation filter 162 as an entropy decoding unit. The point acquired from 191 is different.
- the entropy decoding unit 191 performs entropy decoding on the encoded signal to obtain quantization coefficients, motion data, and filter coefficients (S21).
- the inverse quantization / inverse transform unit 121 inversely quantizes the quantization coefficient and performs inverse DCT transform to generate a quantized prediction error signal (S22).
- the adder 135 adds the quantized prediction error signal and the prediction block to generate a reconstructed signal.
- the deblocking filter 137 removes block distortion from the reconstructed signal and generates a decoded block (S23).
- the image decoding apparatus 101 shown in FIG. 2A applies the post filter 139 to the decoded block to generate an output signal.
- the prediction block generation unit generates a prediction block (S15). Details of the prediction block generation processing will be described later.
- the example of the interpolation filter 162 that operates by obtaining the filter coefficient from the entropy decoding unit 191 is shown, but the present invention is not limited to this. That is, the interpolation filter 162 of both the image encoding device 100 and the image decoding device 101 may determine the filter coefficient by the same method. This eliminates the need to include filter coefficients in the encoded signal, thereby improving encoding efficiency.
- FIG. 3A is a block diagram of the interpolation filter 162.
- FIG. 3B is a flowchart showing a prediction block generation process.
- the interpolation filter 162 includes a pixel selection unit 162a, a weighting factor determination unit 162b, and an interpolation unit 162c. This interpolation filter 162 spatially interpolates the pixels of the image stored in the memory 140.
- the pixel selection unit 162a selects a plurality of pixels used for interpolation of the interpolation target pixel from the image stored in the memory 140.
- the weighting factor determination unit 162b determines a weighting factor corresponding to each of the plurality of pixels selected by the pixel selection unit 162a.
- the interpolation unit 162c calculates the pixel value of the interpolation target pixel by weighting and adding the pixel values of the plurality of pixels using the weighting factor determined by the weighting factor determination unit 162b.
- the prediction block generation unit determines whether to predict a prediction block corresponding to an encoding target block to be encoded next in “inter” mode or “intra” mode, and switches the switch 180. (S31).
- the intra-frame prediction unit 150 generates a prediction block using the decoded signal read from the memory 140 (S32).
- the interpolation filter 162 interpolates the pixels of the image prior to the motion compensation prediction.
- the pixel selection unit 162a selects a plurality of pixels used for interpolation of the interpolation target pixel (S33). For example, a total of six pixels including three integer pixels on the right side of the interpolation target pixel and three integer pixels on the left side are selected.
- the weighting factor determination unit 162b determines a weighting factor for each pixel selected by the pixel selection unit 162a (S34).
- the weighting coefficient corresponding to the above example is ( ⁇ 2, ⁇ 1, 0, 1, 2, 3).
- the weighting coefficient has a different value depending on whether each pixel selected by the pixel selection unit 162a belongs to a target block to which the interpolation target pixel belongs or a peripheral block different from the target block. Specifically, when each pixel belongs to the target block, a larger weight coefficient is assigned than when each pixel belongs to the peripheral block. For example, it is assumed that the above example ( ⁇ 2, ⁇ 1, 0, 1, 2, 3) is a weighting coefficient used when all six pixels belong to the target block. On the other hand, when only the rightmost pixel belongs to the peripheral block, for example, a weighting factor such as ( ⁇ 2, ⁇ 1, 0, 1, 2, 2) is assigned.
- the specific method for determining whether or not each pixel belongs to the target block in the weighting factor determination unit 162b is not particularly limited. For example, first, the distance (horizontal direction and / or vertical direction) from the interpolation target pixel to the boundary of the target block is calculated. Then, for each pixel, whether or not the pixel belongs to the target block may be determined based on whether or not a block boundary of the target block exists between the interpolation target pixel and the pixel.
- the interpolation unit 162c weights and adds the pixel values of each pixel selected by the pixel selection unit 162a using the weighting factor determined by the weighting factor determination unit 162b (S35). Thereby, the pixel value of the interpolation target pixel is calculated. Note that the processing of S33 to S35 is executed for all the interpolation target pixels.
- the motion compensation prediction unit 160 generates a prediction block using the interpolated image and the motion data generated by the motion prediction unit 170 (S36).
- the weight of the pixel belonging to the same block as the interpolation target pixel is relatively increased, and the weight of the pixel belonging to a block different from the interpolation target pixel is relatively decreased. It is. Thereby, the influence of the distortion resulting from encoding for every block can be minimized. As a result, a more accurate prediction block can be generated.
- the value of the weighting factor of the pixels belonging to the peripheral block may be changed according to the boundary strength bS of the block boundary between the target block and the peripheral block. Specifically, the weighting factor may be reduced as the boundary strength bS shown in Table 1 is stronger (the numerical value is larger).
- a smaller weighting factor may be assigned in the former case.
- both the target block and the peripheral block are encoded in the screen
- both the target block and the peripheral block are encoded.
- a smaller weighting factor may be assigned in the former case.
- the target block and the peripheral block do not include coding coefficients
- a smaller weight coefficient may be assigned in the former case.
- the weighting coefficient of the pixel belonging to the peripheral block is changed according to the gradient of the pixel value of two pixels sandwiching the boundary between the target block and the peripheral block among the pixels selected by the pixel selection unit 162a. May be. Specifically, the greater the gradient, the smaller the weight coefficient.
- FIG. 4 is a schematic diagram of a block of video data.
- a black circle indicates the position of a full pixel, that is, a sample position of the original image, and a white circle indicates the position of a sub-pixel whose pixel value is to be interpolated.
- the accuracy of 1/4 pixel is shown.
- the present invention is not limited as long as the accuracy is decimal pixel, and it is not limited to 1/2 pixel, 1/4 pixel, or 1/8 pixel. Can be applied even to those having decimal pixel precision different in the vertical direction and the horizontal direction.
- broken lines 310 and 311 and broken lines 312 and 313 respectively indicate a horizontal block boundary and a vertical block boundary, that is, a 4 ⁇ 4 block boundary.
- a 4 ⁇ 4 block boundary For example, H.I. If it is divided into separate macroblocks according to the H.264 standard, 4 ⁇ 4, 8 ⁇ 4, 4 ⁇ 8, 8 ⁇ 2, 2 ⁇ 8, 8 ⁇ 8, 16 ⁇ 16, 32 ⁇ 16, Regardless of the block size, such as 16 ⁇ 32, 32 ⁇ 32, 64 ⁇ 32, 32 ⁇ 64, 64 ⁇ 64, etc., even a combination of blocks of different sizes within one frame Applicable.
- An original image (that is, a block of video data) is indicated by S x, y which is a pixel value at the full pixel position (x, y).
- An image obtained by interpolating the original image in the horizontal direction and the vertical direction is represented by S ⁇ nx + p, ny + q .
- the symbols “ ⁇ (hat)” respectively indicate the symbols attached immediately above the character, and in the present specification, the symbols “ ⁇ (hat)” are used in the same meaning hereinafter.
- the conventional adaptive interpolation filter is individually defined for each sub-pixel position (p, q). Therefore, such a conventional adaptive interpolation filter can be defined by a linear operator f (p, q) that specifies the position of the original image with respect to an image obtained by interpolating the original image in the horizontal direction and the vertical direction. That is, Formula 1 is materialized.
- Filter coefficient f (i p,, j q ) a set of points that are not in the 0 (i, j), for example, the set of 6 ⁇ 6 pixels that affect the 6 ⁇ 6 taps filter result, usually referred as filter support Is done. Therefore, the summation in Equation 1 and Equation 2 below will exceed the support of the filter used.
- the conventional adaptive interpolation filter is individually defined for each sub-pixel position, but interpolates the pixel value at the sub-pixel position at the center of the block (that is, all pixels used for interpolation belong to the target block).
- the same interpolation filter is used to interpolate pixel values at sub-pixel positions near the boundary of the block (that is, some of the pixels used for interpolation belong to the peripheral block; the same applies hereinafter). use.
- this may be inconvenient in block-based video codecs, especially if there are noticeable blocking artifacts.
- the present invention uses different interpolation filters, for example, when interpolating pixel values near the block boundary and interpolating pixel values in the center of the block.
- the interpolation filter include filters having different filter coefficients, or filters having different filter lengths such as 8 taps, 6 taps, 4 taps, 2 taps, and 0 taps.
- the above description can be realized not only when a plurality of interpolation filters are physically provided, but also by using an interpolation filter whose filter coefficients can be arbitrarily changed.
- the filter length can be substantially changed by setting the weighting coefficient corresponding to the pixels belonging to the peripheral blocks to 0.
- the interpolation filter according to the present invention can be defined with respect to an image obtained by interpolating the original image in the horizontal direction and the vertical direction by a linear operator f (x, y, p, q) that specifies the position of the original image. That is, Formula 2 is materialized.
- f (x, i, y, j p, q) is a discrete filter coefficient (a set of weighting coefficients ) of the interpolation filter, and for x, y, p, q, the current position of sub-pixel accuracy ( nx + p, ny + q).
- the central part of the block is defined as, for example, a part where the filter tap does not contact the block boundary. Furthermore, the same interpolation filter is used for the positions corresponding to different blocks. For example, Expression 4 is established.
- FIG. 5 shows an example of assigning such an interpolation filter to different positions within a block.
- an interpolation filter is selected based on, for example, whether there is a block boundary within a certain distance from the interpolation position.
- This distance is preferably set at half the number of horizontal or vertical filter taps. Therefore, this distance is set to 3 in the 6 ⁇ 6 tap filter.
- the interpolation filter is selected based on whether or not it touches the block boundary.
- the distance is not limited to this distance, and a distance shorter than half the number of filter taps in the horizontal direction or the vertical direction may be employed.
- a pixel value that is longer than half of the number of filter taps in the horizontal direction or the vertical direction from the adjacent block boundary “feels” the block boundary.
- Interpolation may be performed with a simple interpolation filter.
- FIG. 6 shows another example of assigning interpolation filters to different positions in the block.
- an interpolation filter may be selected. Furthermore, since the interpolation filter is generally based on the sub-pixel position in any case, the distance (d 1 , d 2 , d 3 , d 4 ) may be obtained only with full pixel accuracy.
- the boundary strength bS (see Table 1 above) may also be used to select the optimal interpolation filter.
- the value of the parameter bS is also included in the index r.
- these pixels may be You may set the interpolation filter which is not considered. Furthermore, it selectively emphasizes the interpolation effect on these pixels depending on whether a particular pixel belongs to the current block or an adjacent block, and whether these two blocks are more faithfully encoded. Also good.
- ⁇ k, l is the Kronecker delta
- the symbol “ ⁇ (vector)” indicates a symbol added on the immediately preceding character, and the symbol “ ⁇ (vector)” is used in the same meaning hereinafter.
- the component V x and the component V y indicate decimal pixel positions.
- the filter coefficients f (r, i, p, j q) of a specific filter set r are set as follows for an arbitrary sub-pixel position (p, q).
- Ask. P x, y indicates a reference image that has already been decoded
- the prediction error energy for this displacement is expressed as in Equation 6.
- x ⁇ x + [v x / n]
- y ⁇ y + [v y / n]
- [. . . ] Indicates the FLOOR operator which is the maximum value among integers smaller than the independent variable (truncation operator) of the operator.
- the symbols “ ⁇ (tilde)” respectively indicate the symbols added immediately above the character, and in the present specification, the symbols “ ⁇ (tilde)” are used in the same meaning hereinafter.
- the sum of x and y is obtained for the range in which the displacement vector is valid and the interpolation filter of the set r is assigned. This range corresponds to a specific area of the macro book from which the displacement vector is obtained.
- the filter coefficient f (r, i, p, j q) is obtained so that the prediction error energy of Equation 6 is minimized.
- any method may be used as long as it is a numerical optimization algorithm known in this field, such as a gradient descent method and an annealing method.
- the optimal filter coefficient is obtained by solving the simultaneous linear equations obtained by obtaining the partial derivative of Equation 6. May be. That is, Expression 7 is established.
- the two-dimensional interpolation filter f (r, p, q) is separable, it may be described as a combination of two one-dimensional filters g (r, p, q) and h (r, p, q). Good. That is, Expression 8 is established.
- the two-dimensional interpolation is considered to be a two-step process. In the first stage, interpolation is performed in the horizontal direction in order to obtain the pixel value at the sub-pixel position in the “full pixel row”.
- the pixel values in the sub-pixel rows are obtained by performing vertical interpolation on the pixel values obtained in the first stage.
- the filter coefficients of g (r, p) and h (r, p, q) can be easily obtained from Equation 6 and / or Equation 7.
- FIGS. 7A to 7G are schematic diagrams showing selection of an interpolation filter based on the position of a block boundary with respect to a filter tap. Interpolation symmetric 1/2 pixel positions are indicated by circles 630, and full pixel positions at which the interpolation filter (here, 6-tap filter) acts are indicated by squares 620.
- FIG. 7A also shows the situation at the center of the block when the filter tap does not cross the block boundary.
- FIG. 7B to FIG. 7F show five points that can be taken by the sub-pixel position to be interpolated and the adjacent block boundary 610.
- FIG. 7G shows the situation when there are two block boundaries in the support of the interpolation filter. This situation is seen when the filter support is large when the block is a small size such as 4 ⁇ 4, ie, a “large” filter where the number of taps exceeds the block edge length.
- a unique interpolation filter is used for each of the seven cases shown in FIGS. 7A to 7G.
- the number of unique interpolation filters increases or decreases depending on the size of the filter, subpixel accuracy, and block size.
- not all cases shown in FIGS. 7A to 7G have to be processed by a unique interpolation filter. If the filters are symmetric, for example, the filters used in FIGS. 7B and 7C may be the same as the filters used in FIGS. 7F and 7E, respectively. 7G may be the same as the filter of FIG. 7B and the filter of FIG. 7F.
- FIGS. 7A to 7G For simplification, only a one-dimensional embodiment is shown in FIGS. 7A to 7G. These one-dimensional embodiments may be applied directly to two one-dimensional interpolation filters g (r, p) and h (r, p, q) in the case of separable interpolation filters, Two-dimensional interpolation filters (non-separable) can be similarly distinguished by the position of the horizontal and / or vertical boundaries relative to the filter support.
- edges / one or more edges There are no edges / one edge / two edges. ⁇ There is no edge at all / There is an edge at the center / There are other edges. There is no edge / There is an edge in the center / There is one edge in a different position / There are two edges. There is no edge at all / There is an edge at the center / Multiple (one) on the right / Multiple (one) on the left / Two edges.
- parameters that specify block boundaries such as boundary strength bS are also possible.
- the parameters need to be saved with the associated reference frame, the one sent in the bitstream to decode the previous frame, and the one used for filter selection, in order to save these encoding parameters New memory may be added.
- the distance from the interpolation target sample to the nearest block boundary with full pixel accuracy is set to d1 to d4, and the interpolation filter index k is set to min (d1, d3) and min (d2, The reference table specified by d4) is used (see FIG. 6). If the block boundary is divided into blocks in the video encoding method, 4 ⁇ 4, 8 ⁇ 4, 4 ⁇ 8, 8 ⁇ 2, 2 ⁇ 8, 8 ⁇ 8, 16 ⁇ 16, and 32 ⁇ 16. , 16x32, 32x32, 64x32, 32x64, 64x64, etc., or any combination of blocks of different sizes within a frame. May be.
- Each filter index k indicates an interpolation filter that is predicted on the encoder side and transmitted to the decoder.
- the interpolation filters may have different filter coefficients, and may have different filter lengths such as 8 taps, 6 taps, 4 taps, 2 taps, 0 taps, and the like.
- such a reference table is defined as follows.
- the adaptive filter coefficient When the adaptive filter coefficient is transmitted, the additional bit rate increases.
- the overall coding gain may be reduced by overhead information.
- the auxiliary information it may be assumed that the statistical properties of the image are symmetric. In this way, when the distances from the full pixel positions to the current sub-pixel position are equal, the filter coefficients are considered equal.
- the adaptation of the filter has a limit on the signal statistical value, which may lead to a loss of gain in coding efficiency.
- a symmetric interpolation filter and an asymmetric interpolation filter By applying a symmetric interpolation filter and an asymmetric interpolation filter, it is possible to control the amount of overhead information added to the bitstream by transmitting filter coefficients. For example, for high-resolution sequences, an asymmetric interpolation filter may be sent to make the filter optimal for signal statistics, while for low-resolution sequences, a symmetric filter is used to reduce the amount of overhead information. Should be applied. Each time the symmetry is used, the corresponding filters at different sub-pixel positions are optimized together. As a result, when the input signal includes aliasing or when motion prediction is inaccurate, prediction efficiency may be reduced. In order to optimize motion compensation prediction accuracy or overhead transmission, the symmetric filter and the asymmetric filter can be switched to each other according to the sub-pixel position.
- the present invention relates to video encoding and video decoding based on motion compensated prediction with decimal pixel accuracy.
- an interpolation filter selected from a plurality of filter candidates based on the distance from the sub-pixel position to the adjacent block boundary is applied.
- the plurality of filter candidates are configured to optimally interpolate pixels having a specific distance to the adjacent block boundary.
- the boundary type or strength may be taken into account when selecting the optimal filter.
- Each of the above devices is specifically a computer system including a microprocessor, ROM, RAM, a hard disk unit, a display unit, a keyboard, a mouse, and the like.
- a computer program is stored in the RAM or the hard disk unit.
- Each device achieves its functions by the microprocessor operating according to the computer program.
- the computer program is configured by combining a plurality of instruction codes indicating instructions for the computer in order to achieve a predetermined function.
- the system LSI is an ultra-multifunctional LSI manufactured by integrating a plurality of component parts on one chip, and specifically, a computer system including a microprocessor, a ROM, a RAM, and the like. is there. A computer program is stored in the RAM. The system LSI achieves its functions by the microprocessor operating according to the computer program.
- the constituent elements constituting each of the above devices may be constituted by an IC card or a single module that can be attached to and detached from each device.
- the IC card or module is a computer system that includes a microprocessor, ROM, RAM, and the like.
- the IC card or the module may include the super multifunctional LSI described above.
- the IC card or the module achieves its functions by the microprocessor operating according to the computer program. This IC card or this module may have tamper resistance.
- the present invention may be the method described above. Further, the present invention may be a computer program that realizes these methods by a computer, or may be a digital signal composed of a computer program.
- the present invention also relates to a computer-readable recording medium capable of reading a computer program or a digital signal, such as a flexible disk, hard disk, CD-ROM, MO, DVD, DVD-ROM, DVD-RAM, BD (Blu-ray Disc), It may be recorded in a semiconductor memory or the like. Further, it may be a digital signal recorded on these recording media.
- a computer-readable recording medium capable of reading a computer program or a digital signal, such as a flexible disk, hard disk, CD-ROM, MO, DVD, DVD-ROM, DVD-RAM, BD (Blu-ray Disc), It may be recorded in a semiconductor memory or the like. Further, it may be a digital signal recorded on these recording media.
- the present invention may transmit a computer program or a digital signal via an electric communication line, a wireless or wired communication line, a network represented by the Internet, a data broadcast, or the like.
- the present invention may be a computer system including a microprocessor and a memory.
- the memory may store the computer program, and the microprocessor may operate according to the computer program.
- program or digital signal may be recorded on a recording medium and transferred, or the program or digital signal may be transferred via a network or the like, and may be implemented by another independent computer system.
- the example of the image encoding device 100 and the image decoding device 101 including the interpolation filter 162 shown in FIG. 3A has been described. It can be established. That is, the interpolation filter 162 mounted in the image encoding device 100 and the image decoding device 101 can be modularized and provided as an interpolation filter device, an interpolation filter method, an interpolation filter program, an interpolation filter integrated circuit, and the like. .
- the present invention is particularly advantageously used for an image encoding method and an image decoding method using motion compensation prediction.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
(a)各映像フレームのデータ圧縮をブロックレベルで行うために、各映像フレームを2次元の画素ブロックに分割する。
(b)各ブロックに時空間予測方法を適用して、予測誤差を空間領域から周波数領域に変換し、係数を算出することによって、時空間映像情報を無相関化する。
(c)算出した係数を量子化することによって、データの全体量を削減する。
(d)量子化した係数および予測パラメータをエントロピー符号化アルゴリズムによって符号化することで、そのデータを圧縮する。
p0,new = (p2 - (p1<< 1) + (p0 + q0 + 1) >> 1) >> 1
q0,new = (q2 - (q1<< 1) + (q0 + p0 + 1) >> 1) >> 1
・エッジが全くない/エッジが1つ以上ある。
・エッジが全くない/エッジが1つある/エッジが2つある。
・エッジが全くない/中央にエッジがある/その他にもエッジが複数ある。
・エッジが全くない/中央にエッジがある/違う位置にエッジが1つある/エッジが2つある。
・エッジが全くない/中央にエッジがある/右側にエッジが複数(1つ)ある/左側にエッジが複数(1つ)ある/エッジが2つある。
・エッジが全くない/中央にエッジがある/中央から2画素右側にエッジが複数ある/中央から1画素右側にエッジが複数ある/中央から2画素左側にエッジが複数ある/中央から1画素左側にエッジが複数ある/エッジが2つある。
・対称フィルタの場合、エッジが全くない/中央にエッジがある/中央から2画素離れたところにエッジが複数ある/中央から1画素離れたところにエッジが複数ある。
なお、本発明を上記実施の形態に基づいて説明してきたが、本発明は、上記の実施の形態に限定されないのはもちろんである。以下のような場合も本発明に含まれる。
101 画像復号装置
110 減算器
120 変換・量子化部
121,130 逆量子化・逆変換部
135 加算器
137 デブロッキングフィルタ
139 ポストフィルタ
140 メモリ
150 画面内フレーム予測部
160 動き補償予測部
162 補間フィルタ
162a 画素選択部
162b 重み係数決定部
162c 補間部
170 動き予測部
180 スイッチ
190 エントロピー符号化部
191 エントロピー復号部
301,302,303 変位ベクトル
310,311,312,313 破線
320 補間対象画素
610 ブロック境界
620 四角
630 円
Claims (14)
- ブロック単位で符号化された画像の画素を補間する補間フィルタリング方法であって、
補間対象画素の補間に用いる複数の画素を前記画像から選択する画素選択ステップと、
前記画素選択ステップで選択された前記複数の画素それぞれに対応する重み係数を決定する重み係数決定ステップと、
前記重み係数決定ステップで決定された重み係数を用いて、前記複数の画素の画素値を重み付け加算して前記補間対象画素の画素値を算出する補間ステップとを含み、
前記重み係数決定ステップでは、前記複数の画素のうちのいずれかの画素が前記補間対象画素の属する対象ブロックと異なる周辺ブロックに属する場合に、当該画素が前記対象ブロックに属する場合より、当該画素の重み係数を小さくする
補間フィルタリング方法。 - 前記重み係数決定ステップでは、
前記補間対象画素から前記対象ブロックの境界までの距離を算出し、
前記複数の画素それぞれについて、算出した距離を用いて前記補間対象画素から当該画素までの間に前記対象ブロックのブロック境界が存在するか否かを判断することによって、当該画素が前記周辺ブロックに属するか否かを判断する
請求項1に記載の補間フィルタリング方法。 - 前記重み係数決定ステップでは、さらに、前記対象ブロックと前記周辺ブロックとの間のブロック境界の境界強度が高くなるほど、前記周辺ブロックに属する画素の重み係数を小さくする
請求項1に記載の補間フィルタリング方法。 - 前記重み係数決定ステップでは、前記対象ブロック及び前記周辺ブロックの少なくとも一方が画面間予測符号化された場合に、前記対象ブロック及び前記周辺ブロックの両方が画面内予測符号化された場合より、前記周辺ブロックに属する画素の重み係数を小さくする
請求項3に記載の補間フィルタリング方法。 - 前記重み係数決定ステップでは、さらに、前記ブロック境界がマクロブロック境界である場合に、前記ブロック境界がマクロブロック境界でない場合より、前記周辺ブロックに属する画素の重み係数を小さくする
請求項4に記載の補間フィルタリング方法。 - 前記重み係数決定ステップでは、さらに、前記対象ブロック及び前記周辺ブロックの少なくとも一方が符号化係数を含む場合に、前記対象ブロック及び前記周辺ブロックの両方が符号化係数を含まない場合より、前記周辺ブロックに属する画素の重み係数を小さくする
請求項4に記載の補間フィルタリング方法。 - 前記重み係数決定ステップでは、さらに、前記対象ブロック及び前記周辺ブロックの参照フレーム、参照フレーム番号、及び動きベクトル値の少なくとも1つが異なる場合に、前記対象ブロック及び前記周辺ブロックの参照フレーム、参照フレーム番号、及び動きベクトルが全て同一である場合より、前記周辺ブロックに属する画素の重み係数を小さくする
請求項6に記載の補間フィルタリング方法。 - 前記重み係数決定ステップでは、さらに、前記複数の画素のうち、前記対象ブロックと前記周辺ブロックとの境界を挟む2つの画素の画素値の勾配が大きいほど、前記周辺ブロックに属する画素の重み係数を小さくする
請求項3に記載の補間フィルタリング方法。 - 複数の符号化対象ブロックを順次符号化する画像符号化方法であって、
前記符号化対象ブロックの予測ブロックを生成する予測ブロック生成ステップと、
前記予測ブロックを用いて、前記符号化対象ブロックを符号化する符号化ステップとを含み、
前記予測ブロック生成ステップは、
補間対象画素の補間に用いる複数の画素を前記画像から選択する画素選択ステップと、
前記画素選択ステップで選択された前記複数の画素それぞれに対応する重み係数を決定する重み係数決定ステップと、
前記重み係数決定ステップで決定された重み係数を用いて、前記複数の画素の画素値を重み付け加算して、前記補間対象画素の画素値を算出する補間ステップと、
画素が補間された前記画像に基づいて前記予測ブロックを生成する予測ステップとを含み、
前記重み係数決定ステップでは、前記画素が前記補間対象画素の属する対象ブロックと異なる周辺ブロックに属する場合に、当該画素が前記対象ブロックに属する場合より、当該画素の重み係数を小さくする
画像符号化方法。 - 前記符号化ステップでは、さらに、前記重み係数を符号化する
請求項9に記載の画像符号化方法。 - 符号化信号を復号して復号ブロックを生成する画像復号方法であって、
前記復号ブロックの予測ブロックを生成する予測ブロック生成ステップと、
前記予測ブロックを用いて、前記符号化信号を復号する復号ステップとを含み、
前記予測ブロック生成ステップは、
補間対象画素の補間に用いる複数の画素を前記画像から選択する画素選択ステップと、
前記画素選択ステップで選択された前記複数の画素それぞれに対応する重み係数を決定する重み係数決定ステップと、
前記重み係数決定ステップで決定された重み係数を用いて、前記複数の画素の画素値を重み付け加算して、前記補間対象画素の画素値を算出する補間ステップと、
画素が補間された前記画像に基づいて前記予測ブロックを生成する予測ステップとを含み、
前記重み係数決定ステップでは、前記画素が前記補間対象画素の属する対象ブロックと異なる周辺ブロックに属する場合に、当該画素が前記対象ブロックに属する場合より、当該画素の重み係数を小さくする
画像復号方法。 - ブロック単位で符号化された画像の画素を補間する補間フィルタリング装置であって、
補間対象画素の補間に用いる複数の画素を前記画像から選択する画素選択部と、
前記画素選択部によって選択された前記複数の画素それぞれに対応する重み係数を決定する重み係数決定部と、
前記重み係数決定部によって決定された重み係数を用いて、前記複数の画素の画素値を重み付け加算して、前記補間対象画素の画素値を算出する補間部とを備え、
前記重み係数決定部は、前記複数の画素のいずれかが前記補間対象画素の属する対象ブロックと異なる周辺ブロックに属する場合に、当該画素が前記対象ブロックに属する場合より、当該画素の重み係数を小さくする
補間フィルタリング装置。 - コンピュータに、ブロック単位で符号化された画像の画素を補間させるプログラムであって、
補間対象画素の補間に用いる複数の画素を前記画像から選択する画素選択ステップと、
前記画素選択ステップで選択された前記複数の画素それぞれに対応する重み係数を決定する重み係数決定ステップと、
前記重み係数決定ステップで決定された重み係数を用いて、前記複数の画素の画素値を重み付け加算して、前記補間対象画素の画素値を算出する補間ステップとを含み、
前記重み係数決定ステップでは、前記複数の画素のいずれかが前記補間対象画素の属する対象ブロックと異なる周辺ブロックに属する場合に、当該画素が前記対象ブロックに属する場合より、当該画素の重み係数を小さくする
プログラム。 - ブロック単位で符号化された画像の画素を補間する集積回路であって、
補間対象画素の補間に用いる複数の画素を前記画像から選択する画素選択部と、
前記画素選択部によって選択された前記複数の画素それぞれに対応する重み係数を決定する重み係数決定部と、
前記重み係数決定部によって決定された重み係数を用いて、前記複数の画素の画素値を重み付け加算して、前記補間対象画素の画素値を算出する補間部とを備え、
前記重み係数決定部は、前記複数の画素のいずれかが前記補間対象画素の属する対象ブロックと異なる周辺ブロックに属する場合に、当該画素が前記対象ブロックに属する場合より、当該画素の重み係数を小さくする
集積回路。
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
BRPI0905082A BRPI0905082A2 (pt) | 2008-08-18 | 2009-08-13 | método de filtragem de interpolação, método de codificação de imagem, método de decodificação de imagem, aparelho de filtragem de interpolação, programa e circuito integrado |
US12/682,944 US20100226432A1 (en) | 2008-08-18 | 2009-08-13 | Interpolation filtering method, image coding method, image decoding method, interpolation filtering apparatus, program, and integrated circuit |
CN200980100738A CN101828400A (zh) | 2008-08-18 | 2009-08-13 | 插补滤波方法、图像编码方法、图像解码方法、插补滤波装置、程序及集成电路 |
EP09808044A EP2317770A1 (en) | 2008-08-18 | 2009-08-13 | Interpolation filtering method, image encoding method, image decoding method, interpolation filtering device, program, and integrated circuit |
JP2010525583A JPWO2010021108A1 (ja) | 2008-08-18 | 2009-08-13 | 補間フィルタリング方法、画像符号化方法、画像復号方法、補間フィルタリング装置、プログラム、及び集積回路 |
MX2010004122A MX2010004122A (es) | 2008-08-18 | 2009-08-13 | Metodo de filtracion por interpolacion, metodo de codificacion de imagenes, metodo de decodificacion de imagenes, aparato de filtracion por interpolacion, programa y circuito integrado. |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP08014651.7 | 2008-08-18 | ||
EP08014651A EP2157799A1 (en) | 2008-08-18 | 2008-08-18 | Interpolation filter with local adaptation based on block edges in the reference frame |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010021108A1 true WO2010021108A1 (ja) | 2010-02-25 |
Family
ID=40251768
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2009/003886 WO2010021108A1 (ja) | 2008-08-18 | 2009-08-13 | 補間フィルタリング方法、画像符号化方法、画像復号方法、補間フィルタリング装置、プログラム、及び集積回路 |
Country Status (9)
Country | Link |
---|---|
US (1) | US20100226432A1 (ja) |
EP (2) | EP2157799A1 (ja) |
JP (1) | JPWO2010021108A1 (ja) |
KR (1) | KR20110042025A (ja) |
CN (1) | CN101828400A (ja) |
BR (1) | BRPI0905082A2 (ja) |
MX (1) | MX2010004122A (ja) |
RU (1) | RU2010115340A (ja) |
WO (1) | WO2010021108A1 (ja) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012008011A1 (ja) * | 2010-07-12 | 2012-01-19 | 株式会社 東芝 | 動画像符号化方法及び復号方法 |
WO2012020800A1 (ja) * | 2010-08-12 | 2012-02-16 | 日本電信電話株式会社 | 映像符号化方法,映像復号方法,映像符号化装置,映像復号装置およびそれらのプログラム |
WO2012020708A1 (ja) * | 2010-08-12 | 2012-02-16 | 日本電信電話株式会社 | 映像符号化方法,映像復号方法,映像符号化装置,映像復号装置およびそれらのプログラム |
WO2012063878A1 (ja) * | 2010-11-10 | 2012-05-18 | ソニー株式会社 | 画像処理装置と画像処理方法 |
WO2013001720A1 (ja) * | 2011-06-30 | 2013-01-03 | 三菱電機株式会社 | 画像符号化装置、画像復号装置、画像符号化方法及び画像復号方法 |
US20130156111A1 (en) * | 2010-07-09 | 2013-06-20 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video using adjustable loop filtering, and method and apparatus for decoding video using adjustable loop filtering |
JP2013534801A (ja) * | 2010-07-20 | 2013-09-05 | シーメンス アクチエンゲゼルシヤフト | 高解像度の参照フレームを用いるビデオコーディング |
CN112911313A (zh) * | 2021-01-14 | 2021-06-04 | 北京博雅慧视智能技术研究院有限公司 | 一种分像素运动估计方法及其电路装置 |
Families Citing this family (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012503416A (ja) * | 2008-09-18 | 2012-02-02 | トムソン ライセンシング | ビデオ画像の刈り取り方法及び装置 |
US20110135011A1 (en) * | 2009-12-04 | 2011-06-09 | Apple Inc. | Adaptive dithering during image processing |
CN102265611A (zh) * | 2010-03-15 | 2011-11-30 | 联发科技(新加坡)私人有限公司 | 在混合视频编码中具有多个滤波器的本地环路内滤波 |
JP2011199396A (ja) * | 2010-03-17 | 2011-10-06 | Ntt Docomo Inc | 動画像予測符号化装置、動画像予測符号化方法、動画像予測符号化プログラム、動画像予測復号装置、動画像予測復号方法、及び動画像予測復号プログラム |
WO2011126759A1 (en) | 2010-04-09 | 2011-10-13 | Sony Corporation | Optimal separable adaptive loop filter |
WO2011128272A2 (en) * | 2010-04-13 | 2011-10-20 | Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. | Hybrid video decoder, hybrid video encoder, data stream |
CN101841722B (zh) * | 2010-06-08 | 2011-08-31 | 上海交通大学 | 滤波边界强度的检测装置的检测方法 |
US9317896B2 (en) * | 2010-07-09 | 2016-04-19 | Samsung Electronics Co., Ltd. | Image interpolation method and apparatus |
US9344716B2 (en) | 2010-09-28 | 2016-05-17 | Samsung Electronics Co., Ltd. | Adaptive filtering method and apparatus |
US10045046B2 (en) * | 2010-12-10 | 2018-08-07 | Qualcomm Incorporated | Adaptive support for interpolating values of sub-pixels for video coding |
US8761245B2 (en) | 2010-12-21 | 2014-06-24 | Intel Corporation | Content adaptive motion compensation filtering for high efficiency video coding |
CN107172424B (zh) | 2011-03-30 | 2020-04-14 | Lg 电子株式会社 | 环路滤波方法及其设备 |
US20120257702A1 (en) * | 2011-04-11 | 2012-10-11 | Matthias Narroschke | Order of deblocking |
CA2772894A1 (en) * | 2011-05-17 | 2012-11-17 | Her Majesty The Queen In Right Of Canada, As Represented By The Ministerof Industry, Through The Communications Research Centre Canada | Image and video encoding and decoding |
JP5649524B2 (ja) * | 2011-06-27 | 2015-01-07 | 日本電信電話株式会社 | 映像符号化方法,装置,映像復号方法,装置およびそれらのプログラム |
US9641866B2 (en) | 2011-08-18 | 2017-05-02 | Qualcomm Incorporated | Applying partition-based filters |
CN102984435B (zh) * | 2011-09-02 | 2016-02-17 | 晨星软件研发(深圳)有限公司 | 影像处理装置、影像处理方法及影像播放*** |
GB2501535A (en) * | 2012-04-26 | 2013-10-30 | Sony Corp | Chrominance Processing in High Efficiency Video Codecs |
WO2014010584A1 (ja) * | 2012-07-09 | 2014-01-16 | 日本電信電話株式会社 | 画像符号化方法、画像復号方法、画像符号化装置、画像復号装置、画像符号化プログラム、画像復号プログラム及び記録媒体 |
CN108111851B (zh) * | 2016-11-25 | 2020-12-22 | 华为技术有限公司 | 一种去块滤波方法及终端 |
US10965955B2 (en) * | 2016-12-22 | 2021-03-30 | Mediatek Inc. | Method and apparatus of motion refinement for video coding |
TWI833248B (zh) * | 2017-04-06 | 2024-02-21 | 美商松下電器(美國)知識產權公司 | 解碼方法及編碼方法 |
TW201842768A (zh) * | 2017-04-06 | 2018-12-01 | 美商松下電器(美國)知識產權公司 | 編碼裝置、解碼裝置、編碼方法及解碼方法 |
TW202344047A (zh) * | 2017-04-06 | 2023-11-01 | 美商松下電器(美國)知識產權公司 | 解碼裝置、編碼裝置及電腦可讀取之非暫時性記憶媒體 |
WO2018225593A1 (ja) * | 2017-06-05 | 2018-12-13 | パナソニック インテレクチュアル プロパティ コーポレーション オブ アメリカ | 符号化装置、復号装置、符号化方法および復号方法 |
JP7418152B2 (ja) * | 2018-12-17 | 2024-01-19 | キヤノン株式会社 | 画像符号化装置、画像符号化方法、画像復号装置、画像復号方法 |
JP7026065B2 (ja) * | 2019-03-12 | 2022-02-25 | Kddi株式会社 | 画像復号装置、画像復号方法及びプログラム |
US11445174B2 (en) * | 2019-05-06 | 2022-09-13 | Tencent America LLC | Method and apparatus for video coding |
CN110944211B (zh) * | 2019-11-15 | 2022-07-08 | 腾讯科技(深圳)有限公司 | 用于帧内预测的插值滤波方法、装置、介质及电子设备 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004007337A (ja) * | 2002-04-25 | 2004-01-08 | Sony Corp | 画像処理装置およびその方法 |
JP2005072824A (ja) * | 2003-08-22 | 2005-03-17 | Victor Co Of Japan Ltd | ブロック歪低減装置、ブロック歪低減方法、及びブロック歪低減用プログラム |
WO2007011189A1 (en) * | 2005-07-21 | 2007-01-25 | Lg Electronics Inc. | Method of encoding and decoding video signals |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5418714A (en) * | 1993-04-08 | 1995-05-23 | Eyesys Laboratories, Inc. | Method and apparatus for variable block size interpolative coding of images |
JP2008536414A (ja) * | 2005-04-13 | 2008-09-04 | ゴットフリート・ヴィルヘルム・ライプニッツ・ウニヴェルジテート・ハノーヴァー | ビデオの拡張された符号化方法及び装置 |
KR100678958B1 (ko) * | 2005-07-29 | 2007-02-06 | 삼성전자주식회사 | 인트라 bl 모드를 고려한 디블록 필터링 방법, 및 상기방법을 이용하는 다 계층 비디오 인코더/디코더 |
-
2008
- 2008-08-18 EP EP08014651A patent/EP2157799A1/en not_active Withdrawn
-
2009
- 2009-08-13 US US12/682,944 patent/US20100226432A1/en not_active Abandoned
- 2009-08-13 RU RU2010115340/07A patent/RU2010115340A/ru not_active Application Discontinuation
- 2009-08-13 JP JP2010525583A patent/JPWO2010021108A1/ja active Pending
- 2009-08-13 WO PCT/JP2009/003886 patent/WO2010021108A1/ja active Application Filing
- 2009-08-13 EP EP09808044A patent/EP2317770A1/en not_active Withdrawn
- 2009-08-13 KR KR1020107007076A patent/KR20110042025A/ko not_active Application Discontinuation
- 2009-08-13 BR BRPI0905082A patent/BRPI0905082A2/pt not_active Application Discontinuation
- 2009-08-13 MX MX2010004122A patent/MX2010004122A/es not_active Application Discontinuation
- 2009-08-13 CN CN200980100738A patent/CN101828400A/zh active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004007337A (ja) * | 2002-04-25 | 2004-01-08 | Sony Corp | 画像処理装置およびその方法 |
JP2005072824A (ja) * | 2003-08-22 | 2005-03-17 | Victor Co Of Japan Ltd | ブロック歪低減装置、ブロック歪低減方法、及びブロック歪低減用プログラム |
WO2007011189A1 (en) * | 2005-07-21 | 2007-01-25 | Lg Electronics Inc. | Method of encoding and decoding video signals |
Non-Patent Citations (1)
Title |
---|
THOMAS WEDI: "Adaptive interpolation filter for motion compensated prediction", PROCEEDINGS OF 2002 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING., vol. 2, - 10 December 2002 (2002-12-10), pages II-509 - II-512, XP008133028 * |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130156111A1 (en) * | 2010-07-09 | 2013-06-20 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video using adjustable loop filtering, and method and apparatus for decoding video using adjustable loop filtering |
JP2013530659A (ja) * | 2010-07-09 | 2013-07-25 | サムスン エレクトロニクス カンパニー リミテッド | 調節可能なループ・フィルタリングを利用したビデオ符号化方法及びその装置、調節可能なループ・フィルタリングを利用したビデオ復号化方法及びその装置 |
WO2012008011A1 (ja) * | 2010-07-12 | 2012-01-19 | 株式会社 東芝 | 動画像符号化方法及び復号方法 |
US9906787B2 (en) | 2010-07-20 | 2018-02-27 | Siemens Aktiengesellschaft | Method and apparatus for encoding and decoding video signal |
JP2013534801A (ja) * | 2010-07-20 | 2013-09-05 | シーメンス アクチエンゲゼルシヤフト | 高解像度の参照フレームを用いるビデオコーディング |
JP2012044239A (ja) * | 2010-08-12 | 2012-03-01 | Nippon Telegr & Teleph Corp <Ntt> | 映像符号化方法,映像復号方法,映像符号化装置,映像復号装置およびそれらのプログラム |
CN103069803A (zh) * | 2010-08-12 | 2013-04-24 | 日本电信电话株式会社 | 视频编码方法、视频解码方法、视频编码装置、视频解码装置及其程序 |
JP2012044238A (ja) * | 2010-08-12 | 2012-03-01 | Nippon Telegr & Teleph Corp <Ntt> | 映像符号化方法,映像復号方法,映像符号化装置,映像復号装置およびそれらのプログラム |
WO2012020708A1 (ja) * | 2010-08-12 | 2012-02-16 | 日本電信電話株式会社 | 映像符号化方法,映像復号方法,映像符号化装置,映像復号装置およびそれらのプログラム |
TWI501629B (zh) * | 2010-08-12 | 2015-09-21 | Nippon Telegraph & Telephone | 影像編碼方法、影像解碼方法、影像編碼裝置、影像解碼裝置及該等之程式 |
CN103069803B (zh) * | 2010-08-12 | 2016-09-28 | 日本电信电话株式会社 | 视频编码方法、视频解码方法、视频编码装置、视频解码装置 |
US9609318B2 (en) | 2010-08-12 | 2017-03-28 | Nippon Telegraph And Telephone Corporation | Video encoding method, video decoding method, video encoding apparatus, video decoding apparatus, and programs thereof |
WO2012020800A1 (ja) * | 2010-08-12 | 2012-02-16 | 日本電信電話株式会社 | 映像符号化方法,映像復号方法,映像符号化装置,映像復号装置およびそれらのプログラム |
WO2012063878A1 (ja) * | 2010-11-10 | 2012-05-18 | ソニー株式会社 | 画像処理装置と画像処理方法 |
WO2013001720A1 (ja) * | 2011-06-30 | 2013-01-03 | 三菱電機株式会社 | 画像符号化装置、画像復号装置、画像符号化方法及び画像復号方法 |
JPWO2013001720A1 (ja) * | 2011-06-30 | 2015-02-23 | 三菱電機株式会社 | 画像符号化装置、画像復号装置、画像符号化方法及び画像復号方法 |
CN112911313A (zh) * | 2021-01-14 | 2021-06-04 | 北京博雅慧视智能技术研究院有限公司 | 一种分像素运动估计方法及其电路装置 |
CN112911313B (zh) * | 2021-01-14 | 2023-06-02 | 北京博雅慧视智能技术研究院有限公司 | 一种分像素运动估计方法及其电路装置 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2010021108A1 (ja) | 2012-01-26 |
CN101828400A (zh) | 2010-09-08 |
US20100226432A1 (en) | 2010-09-09 |
RU2010115340A (ru) | 2012-09-27 |
BRPI0905082A2 (pt) | 2015-11-17 |
EP2317770A1 (en) | 2011-05-04 |
EP2157799A1 (en) | 2010-02-24 |
KR20110042025A (ko) | 2011-04-22 |
MX2010004122A (es) | 2010-05-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2010021108A1 (ja) | 補間フィルタリング方法、画像符号化方法、画像復号方法、補間フィルタリング装置、プログラム、及び集積回路 | |
RU2694012C1 (ru) | Усовершенствованное кодирование с внутрикадровым предсказанием с использованием планарных представлений | |
JP6685208B2 (ja) | 映像復号化方法及び装置 | |
US9204166B2 (en) | Method and apparatus for decoding intra prediction mode | |
KR101446286B1 (ko) | 쿼드 트리를 이용한 영상 복호화 방법 및 컴퓨터로 판독 가능한 저장매체 | |
KR101947657B1 (ko) | 인트라 예측 정보 부호화 방법 및 장치 | |
CN104967851B (zh) | 用于运动补偿预测的方法和设备 | |
KR101802375B1 (ko) | 시간적 후보 움직임 벡터 유도 방법 및 이러한 방법을 사용하는 장치 | |
JP7525178B2 (ja) | 適応的動きベクトル解像度を用いるビデオ信号処理方法及び装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200980100738.2 Country of ref document: CN |
|
ENP | Entry into the national phase |
Ref document number: 20107007076 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2009808044 Country of ref document: EP |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09808044 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12682944 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: MX/A/2010/004122 Country of ref document: MX |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010525583 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: PI 2010001740 Country of ref document: MY |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010115340 Country of ref document: RU |
|
ENP | Entry into the national phase |
Ref document number: PI0905082 Country of ref document: BR Kind code of ref document: A2 Effective date: 20100416 |